Date of Award
2025
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Mathematical Sciences
Committee Chair
Satyaki Roy
Committee Member
Dongsheng Wu
Committee Member
Summer Atkins
Committee Member
Jennifer Bail
Research Advisor
Satyaki Roy
Subject(s)
Medical care--Decision making--Mathematical models, Hospitals--Administration--Data processing, Reinforcement learning
Abstract
This thesis addresses the challenge of optimizing patient referrals in healthcare systems by integrating clinical needs, geographic proximity, and dynamic infection risk. Healthcare referral networks (HRNs), which capture patient transfers between hospitals, serve as the foundation of the study. First, a recommendation algorithm is developed to assign patients by jointly considering clinical compatibility and logistical considerations like travel distance, striking a balance between quality and accessibility. Building on this, a reinforcement learning framework is proposed to dynamically adjust referral strategies for vulnerable patients by incorporating the evolving risk of hospital-acquired infections. Finally, long-term planning is explored through methods that recommend future hospital placements based on projected population demand and referral patterns. These approaches are validated on real-world HRN datasets using metrics of clinical match, efficiency, and infection-aware allocation. Overall, they open up a data-driven route to resilient, equitable, and adaptive referral systems.
Recommended Citation
Moody, Mitchell, "Multi-factor mathematical optimization for healthcare referral and facility placement" (2025). Theses. 783.
https://louis.uah.edu/uah-theses/783